Practice patterns in both peritoneal dialysis and hemodialysis have changed dramatically in recent years. Most recent clinical studies comparing peritoneal dialysis and hemodialysis have used mortality as outcome. Surprisingly few studies have compared hospitalization or other measures of morbidity. Several methodological issues cloud comparisons, including non-constant hazards ratios, a much higher rate of transfer from peritoneal dialysis to hemodialysis than vice versa, and the very real possibility of differences in ascertainment of morbidity data, such as hospitalization. We examined a prospective, multicenter cohort of 822 consecutive patients starting dialysis therapy. Comorbidity assessment was extensive. Adjusted mortality estimates were similar in hemodialysis and peritoneal dialysis patients. Conclusions about hospitalization rates were heavily dependent on 2 factors: whether comparisons began at inception of dialysis therapy or at 3 months, and whether intention-to-treat or treatment-received analysis was used. Taken together, when early treatment switches were accounted for, peritoneal dialysis was associated with higher hospitalization rates. We also examined the reason for hospitalization. A differential hospitalization pattern was observed, with peritoneal dialysis patients showing higher rates for vascular admissions and admissions related to dialysis technique, and lower rates of non-technique-related infections. Comparison of morbidity is an important, understudied, aspect of hemodialysis/peritoneal dialysis comparisons. Methodological issues can have a large impact on morbidity comparisons. Potentially large morbidity differences may exist without apparent differences in mortality.